METHOD OF OPTIMIZING THE PRODUCTIVITY AND SECURITY OF REAL-TIME CYBERPHYSICAL SYSTEMS BASED ON BALANCE OF TASKS AND RESOURCES

Abstract

The paper investigates the problem of ensuring the security of real-time cyber-physical systems (RFCs) that operate in environments with tight time constraints and are often implemented in programming languages ​​without built-in memory protection, such as C/C++. Traditional security mechanisms designed for general-purpose computing systems create significant additional costs and limit performance, which makes their application in RF CFS problematic. The main goal of the work is to increase the level of security of these systems with minimal impact on the time characteristics of task performance. To achieve this goal, adaptive mechanisms for ensuring the integrity of data flows, control flow, and pointers using conservative estimates of worst-case execution time (WCET) and slack time are proposed. Reserve time is used to dynamically perform additional checks without violating deadlines, which allows you to formally guarantee the timely operation of the system. Data flows and control flow are formalized as graphs with adjacency and reachability matrices, which provides an accurate assessment of the compliance of actual operations with predefined security policies. For pointers, a context-dependent selective check is applied, which reduces the time spent on control. A generalized optimization model has been developed, which allows balancing the level of security and productivity by using the available time reserve and adaptive scheduling of tasks. The proposed step-by-step method includes system modeling, analysis of protection mechanisms, evaluation and use of slack time, adaptive task scheduling, integration of spatial and temporal isolation, dynamic security management, and testing with validation. Empirical studies confirm the effectiveness of the approach: it ensures timely execution of critical tasks, minimizes additional costs for inspections, increases the level of security without compromising productivity. The obtained models and methods create a scientific basis for the further development of adaptive safety mechanisms of the RF CFS and their integration into resource-efficient real-time systems.